Talk:Source code

software anomalies

Hello, my addition to the "see also"-link: Anomaly_in_software was removed on 24th December.

Please rethink, because I think this would be appropriate, reason: anomalies/bugs are also often in sourcecode besides being in docs and somewhere else. I give in the article different examples like: "data flow anomaly" and "control flow anomaly". I would appreciate any discussion on the matter. Thx, ----Erkan Yilmaz (evaluate me!, discussion) 15:32, 27 December 2006 (UTC)

Please tell me if my software is updated please cause some funtions doesn't work Elizabeth Magdalena Miccadei (talk) 07:42, 20 December 2018 (UTC)

Human-readable

I've seen that Rp removed here the assertion that source code is human-readable. How is it debatable that source code "can be naturally read by humans"? The moment it can't be naturally read, it ceases to be source code but an intermediate representation. Also human-readable is opposed to machine-readable, which is explicitly understood as not-source-code. Diego (talk) 10:15, 9 March 2012 (UTC)

This is debatable because one person's intermediate form may be another person's source code (e.g. some people program in assembly language, others only use it as an intermediate form); most Lisp is compiled to a different form, but dedicated Lisp machines have been built; etc.) and all programming languages must be learnt, to a highly varying extent, before they can be read by humans. Furthermore, source code is machine-interpretable - it's just that in most (but not all) cases, the machine does so by first interpreting or compiling the code to a form it can interpret directly.

I removed the phrase specifically because the existing formulation "human-readable computer programming language." suggests that

• All source code is in a programming language (which is not true).
• There are also programming languages which are not human-readable (which is debatable) and text written in those is not called source code (which is also debatable).

and I didn't see a good way to fix those problems. It would be better to fix them. I do agree that source code is code written in a language designed to be more human-readable than, say, code in binary form as interpreted by machines. Rp (talk) 15:14, 9 March 2012 (UTC)

Rp what you say about each person's knowledge is true, but unrelated to the topic. What you describe are high and low-level programming languages, which are a different beast. The source/machine code distinction has always been about the initial form in which the programmer works at whatever the abstraction level.

Even for machine code there is a distinction between Assembly language, which is the text that the programmer edits, and the binary code created by the assembler. When the programmer enter a memory dump in Hex to tweak individual memory cells, that is not a programming language nor considered "the source".

The moment the text is detached from its symbolic human-readable representation and parsed by the interpreter/compiler, it's transformed into an abstract syntax tree that is no longer human-readable and usually only exists in memory; and then it's no source code anymore. Even in Lisp where the language and the program and the AST share the same structure there is a difference between the symbols as sequences as characters in source code and the symbols as references in the compiled memory structure. Some processes can mud the waters a bit (think Java to bytecode, bytecode to a JIT binary, to micro-ops, to transistor signals); but only the first one is the source.

In short, I find that the current version addresses your concerns and it's accurate. It's debatable that there are programming languages that are not human-readable (I can think of Jacquard looms and music boxes), but in that case we couldn't say that they have source code. Diego (talk) 16:09, 9 March 2012 (UTC)

It's not accurate: e.g. the untrue assertion that all source code consists of instructions, statements or declarations has crept back in. Rp (talk) 16:16, 9 March 2012 (UTC)

Good call - though the lead doesn't say that "all" source code is instructions, mentioning comments is sensible. Diego (talk) 16:37, 9 March 2012 (UTC)

(Why there isn't a section for commentary, BTW)? Diego (talk) 16:39, 9 March 2012 (UTC)

The issue is not just comments, but also the fact that many languages are not imperative programming languages. E.g. HTML elements, Prolog clauses, INI file entries, etc. etc. are not instructions; they may be called declarations, but it is not customary to do so. Rp (talk) 13:27, 12 March 2012 (UTC)

Ah, I understand your concern. However, I think as long as some text is called source "code", that's because it's seen as "instructions to a computer" even if those instructions are declarative in nature. Your HTML example would be "instrutions to display a rendered web page in a browser". This uses the meaning of "instructions" in English, not the technical jargon meaning to "instructions available in a processor's instructions set or a programming language's API". Diego (talk) 16:20, 12 March 2012 (UTC)

The act of instructing, teaching, or furnishing with information or knowledge: certainly there's furnishing of information, but what is the act? Rp (talk) 21:33, 12 March 2012 (UTC)

Je suis très content de vous réservoir. Daphis Wesley (talk) 17:45, 19 April 2017 (UTC)

Je suis très content de vous réservoir. Daphis Wesley (talk) 17:45, 19 April 2017 (UTC)

Recent edits

@Timhowardriley The rewrite of the first paragraph was less than good.

The text "source code is a plain text computer program" is incorrect. Source code does not imply a program. One writes a program in source code, but one can write source code without writing a program. Also, saying "A programmer writes the human readable source code" is not wrong, but is awkward and begs the question: who writes the non-human readable code. You are trying to include the notion that source code is human readable and that it's written by a programmer (a human), but the two things are awkward to say together like that.

"The translation process can be implemented three ways" is wrong. There are three described but there are an infinite number of ways to do it.

You italicized "source code" several times that seem out of place. Further you were not consistent with it.

Your comment "Removed immaterial information from intro" is contentious. I think the info you removed was notable. Maybe didn't belong in the intro. But, certainly not immaterial.

I think you struggle with technical writing and writing in general. I think the original text could be written better, but unfortunately, you made it worse.

We have spoken before and you were argumentative I recall. I'm guessing you will be the same now. Stevebroshar (talk) 12:26, 31 July 2024 (UTC)

I think that starting off a discussion with personal attacks is a terrible way to work with others to try to reach a consensus. MrOllie (talk) 12:33, 31 July 2024 (UTC)

Wiki99 summary

Summary of changes as a result of the Wiki99 project (before, after, diff):

• Large-scale rewrite from reliable sources, fixing unsourced content issues
• Added background section to explain the origin and purpose of source code
• Added quality section to cover how source code quality is evaluated and how it is useful to software developers
• Split purposes section to add coverage of important subtopics, such as use of source code for cost estimation, communication, and modification of software
• Improve the section on copyright and licensing based on reliable sources to be less US-centric

Further possibilities for improvement:

• Expand article with more information
• Get the article to good article status

Buidhe paid (talk) 05:51, 5 August 2024 (UTC)